NomadicML raises $8.4M to tackle autonomous vehicle data overload
Transforming AV Data Chaos into Actionable Insights
NomadicML has raised $8.4 million in a seed funding round to address the overwhelming data generated by autonomous vehicles, robots, and other physical AI systems.
The round, at a $50 million post-money valuation, was led by TQ Ventures with participation from Pear VC and Jeff Dean.
NomadicML’s platform employs vision-language models to convert raw video footage into structured, searchable datasets. This approach targets rare edge cases — critical for AI training — that are often buried within massive archives, where roughly 95% of fleet data currently remains unused.
By enabling teams to quickly identify specific scenarios — such as unusual driving conditions or rare operational failures — the platform helps improve fleet monitoring, model training, and iteration speed.
Founders' Journey and Early Success
The company is led by CEO Mustafa Bal and CTO Varun Krishnan, who met as computer science students at Harvard and later worked at companies such as Lyft and Snowflake.
Frustrated by repeatedly encountering the same challenges in managing and extracting value from large-scale video data, the pair founded NomadicML to streamline the process.
Customers including Zoox, Mitsubishi Electric, Natix Network, and Zendar are already using the platform to develop and scale intelligent systems.
Industry users say the tool enables significantly faster scaling compared to manual review or outsourced data labeling processes, while also offering deeper domain-specific insights.
Bal said the platform provides actionable intelligence from proprietary footage, helping autonomous system developers focus on improving performance rather than managing data pipelines.
Future Horizons and Industry Impact
The funding will support customer expansion and continued development of the platform, including efforts to extend capabilities beyond video into non-visual data such as lidar and multi-modal sensor integration.
NomadicML is emerging within a competitive landscape that includes companies like Scale, Kognic, and Encord, as well as Nvidia’s open-source Alpamayo models.
Krishnan described the platform as an “agentic reasoning system,” capable of interpreting complex real-world scenarios by combining multiple models to understand context and actions within video data.
As autonomous systems continue to scale, tools that transform raw data into usable insights are becoming critical infrastructure. NomadicML’s approach highlights a broader shift in the industry — enabling developers to spend less time on data wrangling and more time advancing robotics and AI capabilities.
Andrew Lee
